The blood glucose concentration is ordinarily maintained within a narrow range by a coordinated set of physiologic processes that collectively form a homeostatic system. Abnormalities of homeostatic regulation probably account for most clinical disorders of glucose metabolism, the most common being diabetes mellitus. Investigations of genetic factors involved in glucose homeostasis have focused, by and large, on quantitative genetic analysis of abnormal phenotypes, generally diabetes. The results of much work provide barely partial insight to the genetic factors important in the etiology and the pathogenesis of diabetes; hypoglycemia has received far less attention. We suppose that one major problem with previous work has been improper focus on phenotype, that is diabetes, which is probably remote from the etiological genes concerned, difficult to define, and age-dependent. We have developed mathematical models of homeostasis; in particular, the response of a metrical trait to perturbation has been studied with respect to the lag time (L) between a perturbing force and the time at which restorative forces react, and the strength of that force (b). We speculate that it is these features, especially the lag time, that are under genetic control. We have demonstrated how, in theory, prolongation of the lag time must result in deviation (even uncontrollable deviation) of the metrical trait (such as blood glucose) from one narrow range of variation to a higher range. We propose studying probands and their close relatives to determine how the lag time (calculated from glucose tolerance test data) agrees with direct studies on blood levels of insulin and glucagon, and how it is inherited. Probands will be selected from among individuals (1) without clinical evidence of disorders of blood glucose, (2) newly diagnosed as having type II diabetes, (3) with maturity-onset diabetes of youth (MODY), and (4) with a variety of mendelian disorders that, for different reasons, result in abnormal glucose homeostasis.